Atempo Miria Software Supports Lustre Distributed Storage to Manage Large Data Volumes

Atempo SAS has announced the availability of its Miria solution with support for Lustre distributed file systems, which are principally used in HPC and big data.

Digital transformation associated with ML workflows is generating amounts of data. In all areas of research, such as meteorology, simulations, energy, life sciences, autonomous cars… data is petabyte or even exabyte-scale and typically stored in distributed file systems such as Lustre.

To provide solutions for such specific needs for large customers and strategic partners on major European or global projects (for example: ExCALIBUR in the UK) where storage and compute requirements are massive, the company has added features to support Lustre file systems.

This includes:

Fastscan module that enables the constant collect of created or modified files
Intelligent detection of deleted files
‘Incremental forever’ backup mode
Capacity to restart volumes from backups (DRP)
Reliable recovery via the company’s Time Navigation functionality

Rising to data explosion challenge
Launched in 2019 and already labelled ‘As used by French Armed Forces’ and ‘France Cybersecurity’, Miria is a dedicated software solution for moving, migrating, protecting and archiving large volumes of unstructured data. Miria’s compatibility with Lustre file systems makes it a solution able to answer the needs of the HPC sector – guaranteeing high performance, security, availability and rapid data access.

“We are constantly looking for ways to respond to the challenges faced by our customer and partner ecosystems who must solve complex storage questions in a globalized and highly competitive world. HPCs have singular performance requirements with data flows often exceeding 100GB/s and volumes in excess of five petabytes and Miria is now the solution for such environments,” said Luc d’Urso, CEO, Atempo.

Resource:
Blog

Read also :
Atempo Miria Certified for Nutanix Files Solution
To offer enterprise class data migration [with our comments]
July 2, 2020 | Press Release